Java之JMX获取应用进程CPU/内存使用率，监控GC回收_如何使用jmx 计算cpu 使用率

JMX是什么？

JMX全称是Java Management Extensions。 Java 管理扩展。 它提供了对Java应用程序和JVM的监控和管理功能。

采集应用进程CPU使用率

static SimpleDateFormat format = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss");
public void monitorCPU() {
        try {
            // 操作系统 MXBean
            OperatingSystemMXBean mxBean = (OperatingSystemMXBean) ManagementFactory.getOperatingSystemMXBean();
            // CPU核心数
            int processors = mxBean.getAvailableProcessors();
            // 运行时 MXBean
            RuntimeMXBean runtimeMXBean = ManagementFactory.getRuntimeMXBean();
            // 基于当前时间, 采集应用运行时长, 单位毫秒, 转换成纳秒,
            long prevUpTime = runtimeMXBean.getUptime() * 1000000;
            // 基于当前时间, 采集应用进程CPU运行时长, 单位毫秒
            long prevProcessCpuTime = mxBean.getProcessCpuTime() / processors;
            while (true) {
                // 每隔一秒采集一次
                TimeUnit.SECONDS.sleep(1);
                // 基于当前时间, 采集应用运行时长
                long processCpuTime = mxBean.getProcessCpuTime() / processors;
                // 基于当前时间, 采集应用进程CPU运行时长
                long upTime = runtimeMXBean.getUptime() * 1000000;
                String date = format.format(new Date());
                // 当前运行时长减去上一次的运行时长, 得到采样间隔
                long upTimeDiff = upTime - prevUpTime;
                // 当前应用进程运行时长减去上一次的运行时长, 得到采样间隔
                long processTimeDiff = processCpuTime - prevProcessCpuTime;
                /**
                 * 计算应用进程CPU使用率,
                 * 计算方式: CPU运行时间 / 总运行时间
                 * 例如: upTimeDiff = 1秒, processTimeDiff = 0.2秒, 则应用进程CPU使用率为 20%
                 *
                 * 也就是: 应用进程CPU使用时长, 在一段时间内的使用占比
                 */
                float processCPUUsage = Math.min((float) (processTimeDiff * 100) / (float) upTimeDiff, 100);

                // 将当前采样时间设置为上一次
                prevUpTime = upTime;
                prevProcessCpuTime = processCpuTime;

                System.out.println(date + " -进程CPU- " + String.format("%.2f", processCPUUsage) + "%");
            }
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44

效果图

第二种方式，需要jdk1.8才支持，而且数据不咋的准确。

OperatingSystemMXBean mxBean = (OperatingSystemMXBean) ManagementFactory.getOperatingSystemMXBean();
System.out.println(" -进程CPU- " + String.format("%.2f", mxBean.getProcessCpuLoad() * 100) + "%");
System.out.println(" -系统CPU- " + String.format("%.2f", mxBean.getSystemCpuLoad() * 100) + "%\n");
1
2
3

采集应用进程内存的使用

static NumberFormat fmtI = new DecimalFormat("###,###", new DecimalFormatSymbols(Locale.ENGLISH));
public void monitorMemory() {
        List<MemoryPoolMXBean> memoryBeans = ManagementFactory.getMemoryPoolMXBeans();
        while (true) {
            for (MemoryPoolMXBean mxBean : memoryBeans) {
                StringBuilder sb = new StringBuilder();
                MemoryUsage usage = mxBean.getUsage();
                sb.append("内存模型: [")
                        .append(mxBean.getType().name())
                        .append("], 内存空间名称: [")
                        .append(mxBean.getName())
                        .append("], 初始化[")
                        .append(bytesToMB(usage.getInit()))
                        .append("], 已使用[")
                        .append(bytesToMB(usage.getUsed()))
                        .append("], 可使用[")
                        .append(bytesToMB(usage.getCommitted()))
                        .append("], 最大[")
                        .append(bytesToMB(usage.getMax()))
                        .append("]");
                System.out.println(sb);
            }
            System.out.println("------------------------------------------------------------------------------------------------------");
            try {
                TimeUnit.SECONDS.sleep(1);
            } catch (Exception e) {
            }
        }
    }
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29

效果图

监控应用进程GC

public void monitorGC() {
        /**
         * 以JDK 1.8默认的GC回收器为例: Parallel Scavenge GC
         * gcMXBeans[0] PS Scavenge     监控新生代垃圾回收
         * gcMXBeans[1] PS MarkSweep    监控老年代/元空间/堆外内存急其他
         */
        List<GarbageCollectorMXBean> gcMXBeans = ManagementFactory.getGarbageCollectorMXBeans();
        GCNotificationListener listener = new GCNotificationListener();
        gcMXBeans.forEach(gcmx -> {
            NotificationEmitter emitter = (NotificationEmitter) gcmx;
            emitter.addNotificationListener(listener, null, null);
        });
    }

    private class GCNotificationListener implements NotificationListener {
        @Override
        public void handleNotification(Notification notification, Object handback) {
            CompositeData cd = (CompositeData) notification.getUserData();
            GarbageCollectionNotificationInfo info = GarbageCollectionNotificationInfo.from(cd);
            GcInfo gcInfo = info.getGcInfo();

            Map<String, MemoryUsage> memoryUsageBeforeGc = gcInfo.getMemoryUsageBeforeGc();
            Map<String, MemoryUsage> memoryUsageAfterGc = gcInfo.getMemoryUsageAfterGc();

            System.out.println("GC前 - 内存情况 " + memoryUsageBeforeGc);
            System.out.println("GC后 - 内存情况 " + memoryUsageAfterGc);

            StringBuilder sb = new StringBuilder();
            sb.append("第[")
                    .append(gcInfo.getId())
                    .append("]次")
                    .append("GC, 耗时: [")
                    .append(gcInfo.getDuration())
                    .append("]毫秒, GC名称: [")
                    .append(info.getGcName())
                    .append("], GC原因: [")
                    .append(info.getGcCause())
                    .append("]");
            System.out.println(sb.toString());
            System.out.println("------------------------------------------------------------------------------------------------------");
        }
    }
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42

效果图